domains / families / science / effect-measures

effect-measures

separate

Computes and interprets treatment effect measures (OR, RR, RD, HR, NNT) with calibrated confidence intervals (Wilson, Newcombe, Miettinen-Nurminen, MOVER, profile likelihood, Bender NNT) and reports marginal vs conditional estimands per FDA 2023 covariate adjustment guidance. Use when reporting treatment effects in confirmatory trials, comparing effect sizes across studies, or constructing forest plots.

When to use

• reporting treatment effects in confirmatory trials
• comparing effect sizes across studies
• constructing forest plots
• computing unadjusted OR, RR, RD from contingency tables
• estimating population-level treatment contrasts
• calculating confidence intervals for effect measures

When not to use

• survival analysis for PH violation
• g-computation/standardisation machinery

Skill page

Source